INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

IN A PREVIOUS STUDY ON ENZYME adaptation in the visceral heat exchangers of bluefin tuna (Thunnus thynnus), we found positive thermal compensation in five of seven metabolic enzymes measured (11). These results pointed to the possibility of compensatory gradients of enzyme concentration in other heterothermic tissues. The white muscle of bluefin tuna is known to exhibit a thermal gradient (from skin to core) that is believed to be comparable in magnitude but more stable (4, 5) than that seen in the visceral heat exchangers in this species (3). Work by Carey and Teal (5) as well as our own measurements suggests that the temperature difference between subcutaneous and core white muscle in T. thynnus is about 10°C in temperate (~18°C) waters and remains relatively stable even during prolonged excursions into cold water (4, 24). Although it could be argued that metabolic fine tuning of visceral heat exchanger tissue in T. thynnus is not vital to the animal's function, a similar argument would be more difficult to make for the white muscle, which makes up about 50% of the body mass in tunas (12) and has been shown to exhibit some of the highest glycolytic enzyme activities ever measured in any tissue (15 and the present study). In addition, if we assume that white muscle power is conserved along the thermal gradient,¹ then this would require some sort of flux-conserving mechanism. For these reasons, we expected to find significant evidence of positive thermal compensation with regard to metabolic enzyme activities along the white muscle thermal gradient in T. thynnus.

The heterothermic white muscle in T. thynnus is also an excellent model for testing biochemical predictions of the theory of symmorphosis. Symmorphosis predicts that all components of linear, processive biological systems (such as the mammalian oxygen delivery pathway) are quantitatively matched to one another so that no one component is more limiting than any other (26). Recently, the symmorphosis framework has been used to generate and test interesting hypotheses about biochemical systems, such as enzyme (18, 25) or transporter (9) pathways.

With respect to enzyme compensation in tissues that exhibit stable and predictable temperature gradients, symmorphosis makes explicit predictions for how flux can be conserved most economically. Assuming that the primary energetic cost of performing glycolysis is the synthesis and maintenance of the enzymes, then the most energetically optimal solution will minimize the amount of excess catalytic capacity that exists at every step in the pathway. Although theories of temperature compensation predict higher concentrations or more efficient versions of enzymes at the cold end of a heterothermic tissue (19, 20), biochemical symmorphosis predicts that adjustments in enzymes should be made as a function of the temperature sensitivity of each enzyme. For example, symmorphosis predicts that an enzyme with a Q₁₀ (enzyme activity at 30°C divided by activity at 20°C) of 3.0 will require more dramatic adjustments in concentration along a heterothermic tissue than one with a Q₁₀ of 1.5. If all enzymes are upregulated in the cold tissue proportionally, regardless of their temperature sensitivity, then this might result in adequate but not optimal thermal compensation, because somewhere along the gradient certain enzymes will be present in excess of what is required of them.

We chose to measure enzymes from the glycolytic pathway in this tissue for two reasons. First, glycolysis is the dominant catabolic pathway in fish white muscle, and as a result the glycolytic enzymes in this tissue are abundant and easy to measure. Second, during times of high glycolytic flux (i.e., bouts of burst swimming), it can be assumed that carbon flux proceeds sequentially through the series of enzymes we measured, with very little shunting of intermediates to other pathways (14). In this way, the enzymes of glycolysis in fish muscle are analogous to the linear sequence of structural elements outlined by Taylor and Weibel (26) in their analysis of the mammalian respiratory system.

To test these predictions of thermal compensation and biochemical symmorphosis, we measured the activity of seven consecutive enzymes of anaerobic glycolysis: aldolase, glyceraldehyde-3-phosphate dehydrogenase (G3PDH), phosphoglycerate kinase, phosphoglycerate mutase (PGM), enolase, pyruvate kinase (PK), and lactate dehydrogenase (LDH), plus one enzyme of aerobic metabolism, citrate synthase (CS), at three muscle depths (subcutaneous, intermediate, and deep), each corresponding to a different in vivo muscle temperature. Measuring CS gave us a glimpse into how enzymes of aerobic metabolism change along the thermal gradient, but because CS is not in series with the glycolytic enzymes and is shared by many pathways it could not be used to evaluate the predictions of symmorphosis. Each enzyme assay was conducted at four temperatures, which allowed us to measure the Arrhenius activation energy (E_a) of each enzyme at each muscle depth. This information was used to assess the thermal sensitivity of each enzyme and whether a differential expression of isozymes as a function of muscle depth is likely.

In short, we tested the hypotheses that 1) enzyme thermal compensation occurs along the heterothermic white muscle of T. thynnus and 2) thermal compensation occurs as predicted by the theory of biochemical symmorphosis, i.e., as a function of the thermal sensitivity of each enzyme.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Tissue collection and storage. White muscle samples were collected from six animals. One animal (fork length 155 cm) was caught with rod and reel on August 9, 1995, in the Great South Channel, off the coast of Cape Cod, MA. The other five (fork lengths 184, 183, 181, 193, and 247) were electro-harpooned in the same area on October 17, 1996. Electrically stunning the animal through the harpoon eliminates the lengthy struggles that otherwise accompany landing of these large and powerful fish. Two-centimeter thick steaks of swimming muscle were dissected from each fish at 45% fork length. Strips of white muscle from skin to core were then dissected out of these cross-sections, taking care to exclude the axial red muscle (Fig. 1). These samples were wrapped in foil and frozen in liquid nitrogen within 15 min of capture. Samples were transported to the University of Guelph on dry ice and transferred to a 80°C freezer. All enzyme measurements were made between December 1996 and January 1997. The order in which the enzymes were measured was randomized so as to minimize any effects of enzyme degradation over time. Enzyme activity for all three muscle positions was measured simultaneously at each assay temperature.

View larger version (31K): [in this window] [in a new window]   Fig. 1.   Schematic drawing of swimming muscle cross-section removed for enzyme and lipid analysis. S, I, and C denote locations of the three muscle depths sampled: subcutaneous, intermediate, and core, respectively. Solid lines represent myosepta, and red muscle is shaded. The gap in the bottom one-half of drawing represents the body cavity.

Enzyme assays. White muscle samples were partially thawed and subsamples were dissected from three positions: 1) subcutaneous, 2) intermediate, and 3) core (Fig. 1). Samples from each muscle depth (400-800 mg) were homogenized using a Polytron PT10 unit (Lucerne, Switzerland) in 1:4 (wt/vol) ice-cold 50 mM imidazole buffer (see Enzyme protocols for buffer pH). The homogenate was centrifuged at 23,400 g for 15 min at 4°C, after which the pellet and lipid fractions were discarded. The supernatant was used for all assays. Enzyme activities were measured on a Hewlett-Packard diode array spectrophotometer, model HP8452A (Mississauga, Ontario, Canada) fitted with a thermostatted cell changer that was maintained at the desired temperature with a Haake D8 circulating water bath (Haake Buchler Instruments, Saddlebrook, NJ). Maximal velocity (V_max) for each enzyme was measured at four temperatures: 15, 20, 25, and 30°C. To avoid problems associated with enzyme degradation over time, each set of homogenates was used for the measurement of only four enzymes. Unless otherwise noted, enzyme protocols were modified from the study of Hochachka et al. (17). The millimolar extinction coefficient () was 6.22, and the reaction was monitored at 340 nm unless otherwise indicated. All enzymes were measured in 50 mM imidazole buffer, and all reactions were initiated by addition of substrate. All enzyme assays were optimized with regard to substrate and cofactor concentrations.

Enzyme protocols. Aldolase: pH 7.4, 0.15 mM NADH, excess triosephosphate isomerase and -glycerophosphate dehydrogenase, 0.50 mM fructose-1,6-bisphosphate (omitted for control); CS:  = 13.6, 412 nm, pH 8.0, 0.25 mM 5,5'-dithio-bis(2-nitrobenzoic acid), 0.020 mM acetyl CoA, 0.1 mM oxaloacetate (omitted for control); enolase: 3.0 mM MgSO₄, 50 mM KCl, 0.15 mM NADH, 2.0 mM ADP, excess PK and LDH, 5.0 mM 2-phosphoglycerate (omitted for control); G3PDH [modified from Smith and Lawrence (21)]: 20 mM Na₂HAsO₄, 3.5 mM NAD⁺, 0.10 mM glyceraldehyde-3-phosphate (omitted for control); LDH: 0.15 mM NADH, 5.0 mM pyruvate (omitted for control); phosphoglycerate kinase: 1.0 mM dithiothreitol, 50 mM KCl, 3.0 mM MgSO₄, 0.15 mM NADH, 1.0 mM ATP, excess G3PDH, 20 mM 3-phosphoglycerate (omitted for control); PGM [modified from Grisolia (13)]:  = 13.6, 240 nm, 3.0 mM MgSO₄, excess (PGM-free) enolase, 15 mM 3-phosphoglycerate; PK: 20 mM MgSO₄, 100 mM KCl, 0.15 mM NADH, 5.0 mM ADP, 0.10 mM fructose-1,6-bisphosphate, excess LDH, 5.0 mM phosphoenolpyruvate (omitted for control).

Calculation of enzyme activity. Enzyme activity was expressed in units per milliliter soluble homogenate fraction (instead of units per gram wet mass) due to the presence of a strong gradient of lipid content, from fatty subcutaneous muscle to relatively leaner muscle near the core (Fig. 2). High lipid variability among samples can skew enzyme activity data if traditional dilution and centrifugation protocols are used, i.e., if tissues are diluted on a weight per volume basis and if only the soluble supernatant fraction is used for all assays. The error introduced by this protocol is best illustrated by a simple example. Consider two pieces of muscle tissue, each identical in all ways, including the quality and quantity of enzymes in the myoplasm. The only difference between them is that one of them is composed of 50% lipid by volume and the other is completely devoid of lipid (0%). When these tissues are diluted in buffer and homogenized, the fatty one will contribute only one-half (ignoring differences in tissue density) the enzymes of the lean one. If enzyme assays are performed using only the soluble supernatant fraction, the lean sample will exhibit enzyme activities about two times higher than the fatty sample, even though the metabolically active portion of the tissues is identical in all ways. Because we were only interested in the activity in the metabolically active portion of the tissue, it was necessary to factor out this confounding interaction between the dilution protocol and lipid content. To do this we measured lipid content in all the tissues and expressed activity as units per milliliter tissue soluble fraction. To convert the data to these units, one must know the volume fractions from the nonlipid component of the tissue as well, i.e., what fraction is soluble and will end up in the homogenate supernatant and what fraction is insoluble and will end up in the pellet after centrifugation. For our purposes, it was acceptable to make an assumption about the relative proportions of these fractions, because only the absolute values of activity are affected by the initial assumption, and the relative activities from lean and fatty tissues are virtually unaffected (based on calculations in which the entire range of possible fraction distributions was explored). For the activities reported in this study, we assumed an equal (50/50) split between the soluble and insoluble fractions from the nonlipid component of the tissues.

View larger version (18K): [in this window] [in a new window]   Fig. 2.   Lipid content (mass%) as a function of muscle depth. Note that a significant (P = 0.0012) lipid gradient exists, with subcutaneous muscle possessing about 10 times more lipid than core muscle. Abbreviations are the same as those described in legend for Fig. 1.

Lipid content. Muscle lipid content was measured in five of the six fish using a gravimetric approach based on extraction techniques described in the studies by Bligh and Dyer (2) and Chaiyawat et al. (6). The lack of lipid data for one of the fish meant that V_max measured as units per milliliter of tissue soluble fraction could not be calculated for these samples. However, V_max data from this fish could still be used in the estimation of Q₁₀ and E_a.

Data analyses. E_a was calculated for each enzyme by constructing Arrhenius plots [log₁₀ of enzyme activity vs. the inverse of the absolute temperature (1/K)] (see Fig. 4). E_a was calculated as 2.3 · R · (slope of the Arrhenius plot), where R is the universal gas constant, 1.987 cal · mol¹ · K¹. The slope was estimated using least-squares regression.

1

1

Chemicals. All chemicals were obtained from Sigma Chemical (St. Louis, MO) and were of the highest purity available.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Effect of muscle depth on lipid content. A significant effect (P < 0.01) of muscle depth on lipid content was found, with the mean lipid content (in g lipid/g tissue) of subcutaneous muscle samples exceeding that for core muscle by a factor of 10 (Fig. 2). The lipid gradient is almost surely related to the endothermic life-style of T. thynnus, because no such gradient is known in any ectothermic fish, and concentrating the lipid near the skin maximizes its effectiveness as an insulating layer. Although this phenomenon was not the concern of this study, it certainly merits further investigation.

Effect of muscle depth on enzyme activity. There was no effect (P > 0.05) of muscle depth on soluble fraction enzyme activity for any of the enzymes measured at a given assay temperature (Fig. 3 and see Fig. 5A). We should note that although the activities expressed in units per gram wet mass (not reported here) can be misleading because of the lipid gradient, when our data are expressed in these units, they are consistent with measurements of tuna white muscle glycolytic enzymes measured in other studies (10, 17) and represent some of the highest glycolytic enzyme activities measured in any tissue (15).

View larger version (25K): [in this window] [in a new window]   Fig. 3.   Mean (n = 5) glycolytic enzyme activity (µmol · min1 · ml soluble fraction1) measured at 4 assay temperatures (15, 20, 25, and 30°C), plotted as a function of muscle depth. There was no significant effect of muscle depth on enzyme activity for any of the enzymes measured. G3PDH, glyceraldehyde-3-phosphate dehydrogenase; PGK, phosphoglycerate kinase; PGM, phosphoglycerate mutase; PK, pyruvate kinase; LDH, lactate dehydrogenase. Error bars are SE.

Effect of muscle depth on E_a. No significant effect of muscle depth on E_a was detected for any of the eight enzymes measured (P > 0.05). Figure 4 shows the Arrhenius plots for all the glycolytic enzymes measured. The plots for each muscle depth are nearly identical for each enzyme, both in elevation (a measure of magnitude of activity) and slope (proportional to the E_a). The E_a for G3PDH is by far the lowest, showing that it is relatively insensitive to the effect of temperature. The E_a (in cal/mol ± SE) for each enzyme at all three muscle depths is reported in Table 1.

View larger version (20K): [in this window] [in a new window]   Fig. 4.   Arrhenius plots for each glycolytic enzyme measured. Each line is an Arrhenius plot for a different muscle depth. The similarity of the lines for each enzyme at all 3 depths (both in elevation and slope) demonstrates that there was no effect of muscle depth on either the quantity or quality of enzymes. Note the shallow slope of the plots for G3PDH, which is indicative of this enzyme's near temperature insensitivity.

ACI. Mean compensation indexes (k_subcut/k_core), Q₁₀ values, and ACI values (± SE) for all enzymes are reported in Fig. 5. Figure 5 demonstrates the point that although there is considerable variability among the Q₁₀ of all the enzymes measured, there is relatively little variability in the compensation indexes (k_subcut/k_core) of the same enzymes. This leads to considerable variability among the ACI values of the enzymes, with the ACI values for G3PDH and CS being significantly greater than most of the other enzymes.

View larger version (28K): [in this window] [in a new window]   Fig. 5.   Mean compensation indexes [activity at cold end of thermal gradient/activity at warm end (ksubcut/kcore)] (A), Q10 values (B), and adjusted compensation index (ACI; C), for each enzyme. Values for ksubcut/kcore were calculated from activities from the 2 extremes of the thermal gradient (subcutaneous activity/core activity) measured at 25°C. Q10 values were calculated using values from assays conducted at 20 and 30°C. Values that share a letter do not differ significantly (Fisher's protected least-significant differences test). Error bars are SE. CS, citrate synthase.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

No evidence for adjustments in enzyme concentration or type. The lack of any evidence for compensatory differences in enzyme activity along the heterothermic white muscle in bluefin tuna contrasts with the results that were obtained in a similar study on the heterothermic visceral heat exchangers in the same species (11). In the heat exchanger study, positive thermal compensation was demonstrated in several enzymes, including CS and PK, neither of which showed any compensatory differences in the present study of the white muscle. In addition, the lack of any differences in E_a as a function of muscle depth for all of the enzymes suggests that there is no compensatory gradient of isozyme expression along the thermal gradient, because thermal isozymes frequently display differences in E_a.

Variability among V_max of glycolytic enzymes. One might think that the simplest prediction of biochemical symmorphosis as it applies to enzyme pathways is that the V_max of each enzyme within a pathway should be the same in a given tissue. Otherwise, the enzymes that exhibit the highest activities will be present in excess, and this would be energetically wasteful. However, this reasoning is faulty because it assumes that in vivo conditions are the same as those used to measure V_max, which is known not to be the case. The large variability in V_max values among glycolytic enzymes from the same tissue (such as those in Fig. 3) can be primarily explained by the fact that many of these enzymes catalyze equilibrium reactions and some of the steps in the pathway are held closer to equilibrium than others (1, 23). The closer a reaction is held to equilibrium, the more enzyme is required to ensure that substrate levels and the responsiveness of the system are conserved in the face of large changes in flux (1). In the present study, we were not concerned with the relative proportions of glycolytic enzymes in a given tissue, but rather how the proportions of those enzymes change along a thermal gradient, if at all, and whether they do so in a way that is related to their temperature sensitivity.

A role for G3PDH. The Arrhenius plot for G3PDH (Fig. 4) demonstrates that this enzyme is almost completely temperature insensitive over the physiological range of temperatures experienced by this tissue. The low Q₁₀ (1.2) and E_a (3,700 cal/mol) of G3PDH are not typical for this enzymeQ₁₀ values for G3PDH from the blue mussel, rabbit, cod, and lobster range from 1.8 to 2.7 (7, 8), with E_a values ranging from 14,500 to 19,000 cal/mol (8). It is possible that the temperature insensitivity of G3PDH is the key to how glycolytic flux is conserved along the thermal gradient in bluefin white muscle. The implication of such a mechanism is that whereas the concentrations of all the other enzymes in glycolysis may be just sufficient, and even "optimal" at the cold end of the gradient, these enzymes are present in excess at the warm end of the gradient, with flux through G3PDH being limiting. This implies that although flux may be conserved, this is not accomplished in a way that is predicted by biochemical symmorphosis. It is also possible that enzyme concentrations are optimal at the warm end of the thermal gradient, with flux not being conserved at the cold end of the gradient. This seems unlikely because it implies differential flux rates and hence contractility within the same myotomal cone during burst swimming, as individual myotomes span the entire thermal gradient.

ACI, a quantitative analysis of thermal adaptation. Biochemical symmorphosis predicts that compensatory adjustments of enzyme concentration along a thermal gradient should be made as a function of the temperature sensitivity of the enzyme. In this way, each enzyme is equally limiting at all points along the gradient.

An upper limit for glycolysis. In a similar study of enzymes along the visceral heat exchangers in this species, we demonstrated compensatory gradients of enzyme concentration along the thermal gradient (11). This raises the question of why none of the enzymes in this study showed similar compensatory adjustments. The answer may lie in the extremely high glycolytic activities in tuna white muscle. The glycolytic activities in the present study are the highest ever measured for any frozen tissue (15). Perhaps selection has pushed this tissue to the upper limit of glycolytic catalytic potential (16), leaving simply no scope for upregulation at the cold end of the gradient.